Replication Data for:''Waveform Inversion of Surface Waves Using Generative Adversarial Network''

We proposed a new deep learning surface-wave waveform inversion framework. Synthetic dataset and field data verified the accuracy, reliability and robustness of this method. We used this advanced inversion framework to invert the surface-wave data collected from an electronics factory in Kansas, USA (Miller et al., 1999; The Leading Edge), and obtained an subsurface velocity model with a length of approximately 50m. This model reveals the locations of the underground low-speed zone, pipelines and bedrock surfaces, providing a basis for a preliminary understanding of the local hydrogeological characteristics.